• Aucun résultat trouvé

Argumentation Frameworks with Higher-Order Attacks: Semantics and Complexity

N/A
N/A
Protected

Academic year: 2021

Partager "Argumentation Frameworks with Higher-Order Attacks: Semantics and Complexity"

Copied!
2
0
0

Texte intégral

(1)

HAL Id: hal-02942437

https://hal.archives-ouvertes.fr/hal-02942437

Submitted on 17 Sep 2020

HAL is a multi-disciplinary open access

archive for the deposit and dissemination of sci-entific research documents, whether they are pub-lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

Argumentation Frameworks with Higher-Order Attacks: Semantics and Complexity

Sylvie Doutre, Mickaël Lafages, Marie-Christine Lagasquie-Schiex

To cite this version:

Sylvie Doutre, Mickaël Lafages, Marie-Christine Lagasquie-Schiex. Argumentation Frameworks with Higher-Order Attacks: Semantics and Complexity. 17th International Conference on Principles of Knowledge Representation and Reasoning, Sep 2020, Rhodes (virtual conference), Greece. 2020. �hal-02942437�

(2)

Argumentation Frameworks with Higher-Order Attacks:

Semantics and Complexity

Sylvie Doutre, Mickaël Lafages, Marie-Christine Lagasquie-Schiex

IRIT, University of Toulouse, Toulouse, France {sylvie.doutre, mickael.lafages, lagasq}@irit.fr

Introduction

Argumentation frameworks (AF) are formalisms to express argumentation

problems. In Dung’s one, they are expressed as directed graph in which nodes represent argument and arrow, attack relations between arguments. Higher-order frameworks, unlike Dung’s one, allow to have attacks over attacks. RAF are such a framework (see Figures 1 and 2). Arguments are here represented by circles and attack relations by squares.

For future algorithm investigations, we adapted the notion of Dung’s AF labellings for RAF. We showed the relation between structures (counterpart of extensions for RAF) and different types of structure labellings.

We studied the complexities of RAF decisions problems and shown that despite the higher expressiveness offered by them, the decision classes stay the same as Dung’s AF.

Figure 1 Figure 2

Examples of RAF

Complexities of RAF

We introduced a new flattening of RAF to Dung’s AF (procedure called Raf2Af) in order to prove that it is also the case for RAF complexities.

Table 3 summarises the complexities of the credulous and skeptical acceptance problems, the verification, the existence, the non-empty existence and the uniqueness problems.

Figure 3 shows an example of flattening. For each attack two arguments are created : one, named as the attack, representing the validity of the attack, the other one the validity of both the attack and its source. For each argument, an other one is created representing the invalidity of the argument.

Same complexities as Dung’s AF

Figure 3 : Raf2AF(Γ) of Figure 2 RAF

Labellings for RAF

Instead of extensions (set of arguments), RAF solutions are expressed as

structure: a couple of sets, one of arguments and one of attacks. As for

Dung’s AF, we introduced structure labellings for RAF, a couple of labellings, one for the arguments and the other one for the attacks. They are three value-based: in (accepted), out (rejected), und (undecidable).

Reinstament RAF labellings are particular labellings that coïncide under

some constraints to differents RAF semantics (see Tables 1 and 2).

Table 1: RAF labellings for Figure 1

Table 3: Complexities for RAF decision problems

Perspectives

• Algorithms for RAF argumentation problems

• Complexities of function problems

Références

Documents relatifs

The connection between abstract argumentation and logics goes back to the seminal work of Dung, where a translation from an AF to a logic program was given. This line of research

This paper proposes a new family of seman- tics which rank-orders arguments from the most acceptable to the weakest one(s). The new semantics enjoy two other main features: i) an

Note that we reduced our problem of giving semantics to higher level networks of the form (S, R) to the problem of giving semantics to the joint attack network (S ∪ R, R) of

This enables us to define two semantics accounting for two requirements: selecting maximal vs-admissible sets and selecting sets offering a strongest defence for their

Despite these efforts in analyzing security protocols, because of the differences between protocols and APIs already noted, verification of security APIs did not begin until

/ La version de cette publication peut être l’une des suivantes : la version prépublication de l’auteur, la version acceptée du manuscrit ou la version de l’éditeur. Access

The paper is organized as follows: the Section 2 contains basic facts on the action of groups of matrices on homogeneous polynomials and Section 3 recalls facts on reflection groups,

1) Managing feedback loops: Up to this point we have constrained our definition of arithmetic operator to functions. In fact, the current implementation of FloPoCo can also